A central mandate of the ADRC is to understand the pathophysiologic processes that underlie dementia. Nonetheless, the physical basis of cognitive and psychiatric symptoms in Alzheimer's disease (AD) remains uncertain, and there is a gap in our knowledge about age related changes in nondemented elderly. The Massachusetts ADRC has been collecting brain tissue from individuals who have been longitudinally clinically studied for 10 years. We have accumulated enough material to finally begin to construct a detailed model of clinical-pathological relationships in AD. We now propose to expand studies we recently initiated aimed at uncovering the temporal sequence of neuropathological changes in AD, and at answering whether or not neurofibrillary tangles (NFT), senile plaques (SP), neuronal loss an/or synaptic loss correlate with the clinical symptoms of AD. We will study AD patients, Down syndrome patients, and nondemented elderly individuals. Our results so far have led to several unexpected findings: SP do not correlate at all with duration or severity of dementia. We hypothesize that SP are in a dynamic equilibrium, both forming and resolving during the course of the disease. We suggest that A beta may be mobilized from the neuropil by ApoE or other potential A beta chaperone proteins. We will explore the recently reported association between sporadic AD and the ApoE-E4 allele by studying the influence of ApoE genotype on neuropathological phenotype. We also postulate that interactions of SP with inflammatory or proteolytic processes could lead to turnover of SP, and potentially to loss of neurons and synapses as well. We will examine SP turnover using a novel strategy based on the nonenzymatic accumulation of advanced glycosylation end products on long lived proteins. We and others find that NFT number in high order association cortices correlate with global measures of dementia. Synaptic loss has also been suggested as a physical substrate of dementia. We find that estimates of neuronal loss in high order association cortex correlate with clinical condition remarkably well. Surprisingly, however, our preliminary data suggest that neuronal loss far outstrips NFT formation, and that NFT formation may account for only a small minority of neuronal loss in the neocortex. Finally, neuropathological variables will be combined with the clinical data base. This will allow us to examine clinical and neuropathological heterogeneity, and to test hypotheses about clinical-pathological correlations. Looking to the future, we expect that additional candidate genes that are risk factors for or causative of AD will emerge, and the tissue samples and data developed in this project will be an ideal resource for determining clinical-pathological and genotype- phenotype correlations.